JP3971171B2 - Copper sputter target processing method - Google Patents

Description

【００２８】
下記の表２は、アニーリング温度が粒度に与える最終作用を例示する。
【００２９】
【表２】

【００３０】
表１及び表２に掲記するプロセスは、純度９９．９９及び９９．９９９パーセントの銅において粒度を調節するための有効な手段を提供した。加えて、このプロセスは、等軸の粒子構造を有しかつ（１１１）、（２００）、（２２０）及び（３１１）配向の各々が５０パーセントよりも少ない銅を生成した。比較サンプルＡ〜Ｄは、熱圧がなく、（２２０）方向で配向される粒子を５０パーセントよりも多く含有していた。
【００３１】
例２
テストは、また、いくつもの連の純度９９．９９９９パーセントの銅スラブを含むものであった。これらの銅スラブは、粒子を束縛し（ｐｉｎ）かつ粒子成長を防ぐ不純物が無いことにより、一層精密な制御を要する。このテストは、直径１５．５６ｃｍ及び厚さ４．９８ｃｍを有する丸い断面のスラブに依拠した。すべてのサンプルを、温度に加熱した炉の中に入れて１時間アニールした。
【００３２】
下記の表３及び表４に示す実験は、アニーリング温度が発明のプロセスに与える作用をテストする。
【００３３】
【表３】

【００３４】
【表４】

【００３５】
【表５】

【００３６】
表５は、アニーリング温度を約４２５℃よりも低く保つと、最適な最終粒度をもたらすことを例示する。その上に、そのプロセスは、単一の又は複数の工程で作動する。
【００３７】
例３
直径１２．７０ｃｍ及び高さ１０．１６ｃｍを有する純度９９．９９９パーセントの銅ビレットを周囲条件下で温度６５０℃に約１時間加熱した。その物質を、次いで周囲条件下で中間アニールしないで６０パーセント熱圧した。熱圧された物質を、次いで空気中で冷却し、続いて冷間圧延して６０パーセント低減させた。ターゲットブランクを、次いで周囲条件下で温度３５０℃において約２時間アニールし、２つの反対の位置の板の間で扁平にし、アルミニウム支持板に爆発結合させた。スパッターターゲット／支持板アセンブリーは、平均粒度約２５μｍ及びアスペクト比１．０５を有するスパッターターゲットを示し、これは、ターゲット加工する間に得られかつ爆発結合プロセスの間保たれた。観測された結晶学的構造は、（１１１）、（２００）、（２２０）及び（３１１）配向の等しい比を示した。
【００３８】
本発明を発明の実施態様を記載することによって例示し、かつ実施態様を相当に詳細に記載したが、特許請求の範囲の記載をそのような詳細に限定する又はいずれかでも制限することを意図しない。更なる利点及び変更は、当業者に容易に明らかになるものと思う。従って、発明は、それの一層広い態様では、示しかつ記載する特定の詳細、代表的な装置及び方法並びに具体例に制限しない。よって、出願人の全体的な発明の概念の範囲又は精神から逸脱しないで、そのような詳細から逸脱してよい。[0001]
BACKGROUND OF THE INVENTION
The present invention provides high purity copper for both reducing particle generation in sputtering processes and increasing film uniformity on substrates such as semiconductor wafers used in the manufacture of semiconductor devices and circuits. The present invention relates to a method for processing a sputter target.
[0002]
[Prior art]
Sputtering refers to a process that involves coating a semiconductor wafer or other substrate mounted within a processing chamber. This chamber contains an inert gas ionized by an electric field and a sputter target spatially opposed to the wafer. The sputter target houses the electrical bias to the wafer. Ions from the gas impact the target, expelling atoms from the target and attaching the target material to the wafer.
[0003]
In the production of sputter targets used in the semiconductor industry, especially sputter targets used in the physical vapor deposition (PVD) of thin films on complex integrated circuits, 1) film uniformity; 2) high deposition rate; 3) It is desirable to produce a sputter target that will have minimal particle generation during spattering; and 4) good conductivity for connecting transistors. For example, sputtering aluminum and aluminum alloy sputter targets deposits a thin conductive film on the integrated circuit for interconnect purposes. However, copper has the potential to replace aluminum in interconnect technology. Copper has both high electrical conductivity and a greater resistance to electromigration compared to aluminum. Other possible advantages of using copper interconnect films include reducing both power loss and interconnect signal delay.
[0004]
As particle size increases and particle size uniformity decreases, target performance decreases. Moreover, it is known that the crystallographic orientation of the sputter target and the distribution of the material ejected from the target affect the film uniformity and sputter deposition rate. It is also known that sputtering of atoms from the target is preferentially performed along the dense direction of the target material and near random particle orientation results in better uniformity of the sputtered film.
[0005]
Conventional processes for producing aluminum or copper targets result in an oriented crystal structure of either (200) or (220). However, a target with a strong (200) or (220) crystal orientation results in a film with poor uniformity. Thus, it is desirable to have a target with random or weak orientation.
[0006]
The copper sputter target can contain second phase alloy precipitates up to 10 μm in size to control particle size. However, the poor conductivity of large second phase precipitations can cause local arcing during sputtering and deposit disadvantageous high density or large particles. In addition, using the second phase to control particle size does not give effective control to a copper target having a purity of 99.99 percent or higher.
[0007]
In conventional target cathode assemblies, a single coupling surface attaches the target to a non-magnetic support plate, typically an aluminum or copper support plate. This forms a parallel interface between the sputter target and the support plate in the assembly. The support plate provides a means for holding the target in the sputtering chamber and provides structural stability to the target. Further, the support plate is cooled with water to remove heat generated by ion bombardment of the target. When the target and support plate are attached by techniques such as soldering, brazing, diffusion bonding, clamping, screw fastening or epoxy bonding, good thermal and electrical contact between the target and the support plate is achieved. . Unfortunately, solder bonding is susceptible to delamination during the sputtering process. In addition, the relatively low bonding temperature associated with “soft” solder lowers the temperature range of the target for sputtering. Thus, soldered assemblies are more expensive and time consuming for consumers, because they are used at lower power levels to avoid separating the target and support plate. Because it must be. This causes a reduction in the sputtering rate.
[0008]
Diffusion bonding, particularly with pretreated rough surfaces, results in a stronger bond. However, preparation for diffusion bonding takes time. More importantly, the high temperatures involved in diffusion bonding change the microstructure obtained during prebonding. Thus, even though fine grain size and random orientation can be achieved during the target manufacturing process, they are lost by current diffusion bonding techniques. Diffusion bonding has the effect of doubling the particle size for pure copper targets. Thus, delamination and changes in microstructural and metallurgical properties are a significant disadvantage of conventional diffusion bonding techniques, making them undesirable for copper target assemblies where small uniform particles are desirable in sputter targets. .
[0009]
An alternative to using a monolithic sputter target without a support plate is also the continuing need for an increase in target diameter to sputter larger silicon wafers and a requirement to increase the purity of the target material Both of these are difficult to implement in view of the increased cost for monolithic targets.
[0010]
[Problems to be solved by the invention]
Thus, there is a need to provide a method for processing a pure copper target assembly in which fine equiaxed uniform particle structures and random crystallographic structures are strongly bonded to a non-magnetic support plate.
[0011]
[Means for Solving the Problems]
Summary of the invention The method is used to fabricate a pure copper sputter target. The method includes first heating the copper billet to a temperature of at least about 500 ° C. The copper billet has a purity of at least 99.99 percent. The copper billet is then hot worked to strain at least about 40 percent. The hot worked copper billet is then cold rolled to add at least about 40 percent strain and form a copper plate. Finally, the copper plate is annealed at a temperature above about 250 ° C. to form a target blank. The target blank has equiaxed particles with an average particle size of less than about 40 μm. The crystallographic structure of the target blank contains (111), (200), (220), and (311) oriented particles, and the amount of target blank particles having each of the orientations is less than about 50 percent.
[0012]
DETAILED DESCRIPTION OF THE INVENTION
DETAILED DESCRIPTION OF THE INVENTION Processing high purity copper by a process that includes heating, hot working, cold working and annealing steps produces sputter targets with randomly oriented fine, uniform particles. To do. This process applies these steps to high purity copper ingots, billets, plates or any other form suitable for subsequent processing operations. High purity copper has a purity of at least 99.99 percent. Advantageously, the copper has a purity of at least 99.999 percent. Most advantageously, the copper has a purity of at least 99.9999 percent to limit the generation of impurity-containing particles. Most advantageously, the sputter target is explosively bonded to the support plate to produce the support plate without changing the size or orientation of the particles.
[0013]
First, high purity copper is heated to a temperature above about 500 ° C. to prepare for hot working the copper. This step preheats the copper for at least 30 minutes to ensure uniform heating. In addition, the thicker the cross-section of the high-purity copper billet, the longer the required preheating. Advantageously, high purity copper is heated to a temperature in the range of about 500 ° C. to 750 ° C. for a period of about 1 to 6 hours to prepare the billet for hot working. High purity copper is heated to a temperature in the range of about 600 ° C. to 700 ° C. for a period of about 1 to 6 hours to improve the final grain size after annealing. Heating the copper billet to a temperature of about 625 ° -675 ° C. produces the most advantageous results. The atmosphere in which the material is preheated is not critical. The material may be heated under ambient conditions or in a protective atmosphere to minimize oxidation.
[0014]
The high purity copper material is then hot worked by applying at least 40 percent strain at a temperature of at least about 500 ° C. to break up the remaining large particles. Because of this specification, hot working defines a temperature that applies strain to impart energy to copper without excessive cracking or excessive grain growth-under these conditions, partial recrystallization or near-complete recrystallization. Crystals can occur. Advantageously, hot working at about 40-80 percent strain, such as by hot pressing, hot forging, hot rolling, equal channel angle extrusion or any other suitable metal working operation, is best. Results in. Hot working may not break the particles well for subsequent cold working steps if the strain is less than 40 percent, and will result in uncontrolled particle growth if the strain exceeds 80 percent. . Hot working imparts about 50-70 percent strain to high purity copper. The amount of processing as used herein is defined as either the ratio between thickness reduction and starting thickness multiplied by 100 percent or the amount of shear strain applied to the material. Hot working may be performed using an intermediate annealing step or without an intermediate annealing step.
[0015]
Hot working will result in cracking if the temperature is below 500 ° C and grain growth if the temperature exceeds 750 ° C. A temperature of about 500 ° -750 ° C. provides effective particle size destruction without undue particle growth. Advantageously, hot working at a temperature of about 600 ° to 700 ° C. allows a wide range of strain to be imparted to the copper. Hot working is most advantageously performed at a temperature of about 625 ° to 675 ° C. Again, the atmosphere in which the material is hot worked is not critical. The hot-processed material is then cooled to about room temperature by air cooling, oven cooling, quenching in water, or quenching in any other medium.
[0016]
The hot worked material is then cold rolled with a strain of at least 40 percent to further break the particles. For this specification, cold rolling is defined as rolling at a temperature with copper at ambient conditions or with copper preheated to a temperature below 250 ° C. Advantageously, cold rolling at about 40-80 percent strain, most preferably about 50-70 percent strain, further breaks the particles to form a target blank of the desired final thickness. Most advantageously, cold rolling causes the copper to have a thickness of less than about 2 cm.
[0017]
Cold rolling with copper preheated to a temperature below 250 ° C. prevents grain growth during cold rolling. Cold rolling is advantageously performed with copper preheated to a temperature below 200 ° C. Cold rolling is most advantageously performed under ambient conditions.
[0018]
Cold rolling is followed by annealing with a target blank at a temperature above about 250 ° C. to obtain a final microstructure of fine, uniform grain size and random orientation. Annealing is preferably performed at a temperature of about 250 ° to 400 ° C, and most preferably at about 300 ° to 375 ° C. If the annealing temperature is too low, the copper will not be recrystallized, and if the temperature is too high, uncontrolled grain growth will occur. As with the heating step, the target may be annealed under ambient conditions or in a protective atmosphere to minimize oxidation of the target material.
[0019]
Advantageously, annealing for at least 30 minutes results in sufficient recrystallization. A typical commercial operation would use an annealing cycle of about 0.5-12 hours. Annealing is most advantageously performed with an annealing cycle of about 1-2 hours.
[0020]
This process produces particles that are smaller than about 40 μm in size. The particles advantageously have a final size of about 10 to 35 μm. Most advantageously, the final particle size is about 10-30 μm.
[0021]
The crystallographic structure of the target blank after annealing contains (111), (200), (220) and (311) oriented particles. The amount of target particles having each of the above orientations is advantageously less than about 50 percent. Most advantageously, the amount of target particles having each of the above orientations is less than about 40 percent of each particle orientation. For example, typical target particles advantageously have about 10 to 40 percent of each of the above particle orientations, and most advantageously about 15 to 35 percent of each of the above particle orientations.
[0022]
After the particles are annealed, the aspect ratio defined herein as the ratio of particle length (cold rolling direction) divided by width (cold rolling transverse direction) is less than about 1.5. And most advantageously less than about 1.3, indicating equiaxed particles. The temperature is raised to avoid further work on the target material, but the top and bottom surfaces of the target blank are flattened between two oppositely positioned plates, but the target thickness is significantly increased. It may be flattened, for example, by pressing at a pressure that does not reduce the pressure.
[0023]
Since the above process results in the desired target microstructure to achieve good uniformity and minimal particle generation during sputtering, further manufacturing steps to complete the target assembly are directed to the microstructure. We must minimize changes. In particular, the technique for bonding the target to the support plate and machining the target to the final dimensions should avoid imposing high temperatures or additional work on the target material.
[0024]
Thus, in accordance with the further principles of the present invention, the target blank is explosively bonded to a copper, aluminum or any other suitable metal or alloy support plate. Suitable metals are those with good thermal conductivity, high strength and elasticity, whose thermal expansion coefficient is comparable to copper. Explosive coupling is used to place the surface of the target in close proximity to the surface of the support plate (ie, reduce the isolation space between them) and accelerate one or both of the surfaces in the direction of each other. This is accomplished by creating one or more managed detonations. The explosive bonding method is described in Paul S. Gilman et al., Fully described in co-pending US patent application Ser. No. 09 / 349,285 entitled “Method of Bonding a Splattering Target to a Baking Plate”, the entirety of which is incorporated herein by reference. Incorporate.
[0025]
This explosive coupling method creates strong atomic / metallurgical bonds in the form of wave-type morphology at the interface between the sputter target and the support plate, but the metallurgical properties of the sputter target or support plate material. No change occurs except for local deformation inherent in the wave-type coupling interface. The explosive bonding process takes place at a non-high temperature, and the heat generated from the process is generated for a period of time that is insufficient for heat to conduct to the support plate and target material; thus, in the target and support plate metal, There is no appreciable temperature rise that produces grain growth. Thus, the finely oriented particles achieved by processing the target blank are not altered by the explosive bonding process and a strong bond is achieved.
[0026]
【Example】
Example 1
The test relied on a square copper slab having a length of 3.81 cm and a thickness of 5.31 cm. Samples A to D represent comparative examples, and samples 1 to 8 represent examples of the invention. Samples AD and Samples 1-8 had a purity of about 99.99 percent, and Samples 9-11 had a purity of at least 99.999 percent.
[0027]
[Table 1]

[0028]
Table 2 below illustrates the final effect that annealing temperature has on particle size.
[0029]
[Table 2]

[0030]
The processes listed in Tables 1 and 2 provided an effective means for adjusting the grain size in 99.99 and 99.999 percent pure copper. In addition, this process produced copper with an equiaxed grain structure and each of (111), (200), (220) and (311) orientations being less than 50 percent. Comparative Samples A-D contained no more than 50 percent particles that were not hot pressed and oriented in the (220) direction.
[0031]
Example 2
The test also included a series of 99.9999 percent purity copper slabs. These copper slabs require more precise control due to the absence of impurities that pin the particles and prevent particle growth. This test relied on a round cross-section slab having a diameter of 15.56 cm and a thickness of 4.98 cm. All samples were annealed for 1 hour in a furnace heated to temperature.
[0032]
The experiments shown in Tables 3 and 4 below test the effect of annealing temperature on the process of the invention.
[0033]
[Table 3]

[0034]
[Table 4]

[0035]
[Table 5]

[0036]
Table 5 illustrates that keeping the annealing temperature below about 425 ° C results in an optimal final particle size. In addition, the process operates in a single or multiple steps.
[0037]
Example 3
A 99.999 percent pure copper billet having a diameter of 12.70 cm and a height of 10.16 cm was heated to a temperature of 650 ° C. for about 1 hour under ambient conditions. The material was then hot pressed 60 percent without intermediate annealing under ambient conditions. The hot pressed material was then cooled in air, followed by cold rolling to reduce 60 percent. The target blank was then annealed under ambient conditions at a temperature of 350 ° C. for about 2 hours, flattened between two oppositely positioned plates and explosively bonded to an aluminum support plate. The sputter target / support plate assembly exhibited a sputter target having an average particle size of about 25 μm and an aspect ratio of 1.05, which was obtained during target processing and maintained during the explosion bonding process. The observed crystallographic structure showed equal ratios of (111), (200), (220) and (311) orientations.
[0038]
Although the invention has been illustrated by describing embodiments of the invention and described in considerable detail, it is intended that the claims be limited to such details or in any way limited do not do. Further advantages and modifications will be readily apparent to those skilled in the art. Accordingly, the invention in its broader aspects is not limited to the specific details, representative apparatus and methods, and specific examples shown and described. Accordingly, departures may be made from such details without departing from the scope or spirit of applicant's general inventive concept.

Claims (1)

following:
Heating a copper billet having a purity of at least 99.999 percent to a temperature of 600-700 ° C;
Hot working a heated copper billet having a temperature of 600-700 ° C. to a strain of 50-70 percent;
Cold rolling a hot-worked copper billet to add 50-70 percent strain and forming a copper plate; and annealing the copper plate at a temperature of 300-375 ° C. to have an average grain size of 10-30 μm And a method of processing a copper sputter target comprising forming a target blank containing (111), (200), (220) and (311) orientations, wherein the amount of particles having each orientation is less than 50 percent.